Table 2

Results of the Multivariate analysis of covariance (MANCOVA) applied to scales A, B and C of the Autism Diagnostic Interview-Revised (ADI-R) and the three univariate regressions separately relating each of the ADI-R scales to the covariatesa

Multivariate analysisUnivariate analysis
TermPillai
statistic
F(3,137) (P)η2ADI-R, scale AADI-R, scale BADI-R, scale C
F(1,139) (P)η2F(1,139) (P)η2F(1,139) (P)η2
Gender0.073.4 (0.019)0.075.7 (0.018)0.049.0 (0.003)0.064.9 (0.029)0.03
Age0.052.2 (0.090)0.055.2 (0.024)0.041.3 (0.253)0.012.4 (0.126)0.02
Verbal IQ0.104.9 (0.003)0.116.8 (0.01)0.0515.0 (<0.001)0.102.1 (0.150)0.01
Group0.073.2 (0.025)0.063.3 (0.070)0.024.0 (0.046)0.038.7 (0.004) 0.06
  • a. Univariate analysis are all based on type III sums of squares. Partial eta-squared (η2) is a measure of effect size (small: 0.01; medium: 0.09; large: 0.25; based on the square of the Pearson correlation effect sizes from Cohen19). Multivariate η2 = 1 – Λ1/s where Λ = Wilks' lambda and s = minimum of the number of levels of the factor minus 1, or the number of dependent variables (here, s = 1 for continuous variables).20